There are several different approaches for implementing auto-focus in a digital camera. According to a relatively simple and inexpensive approach, when a user presses a shutter button part-way down, the camera takes a series of exposures while moving the lens through a focus range. An auto-focus region at the center of the sensor is read out, and a focus metric (a value identifying the extent to which objects are in focus) is generated. Logic in the camera then identifies the focus metric corresponding to the best focus position for the camera lens. However, this technique is slow, since the lens must be stepped through a number of possible focus positions. As a result, there is a relatively long delay period (e.g., about one second) between the time that the shutter button is pressed and the time that a picture is taken.
Another approach for implementing auto-focus in a camera involves using an auto-focus module comprising optical elements such as prisms or beam-splitters, in addition to an array of light-sensing elements. This approach has the advantage of rapidly measuring a desired focus distance based on input from the optical elements. As a result, the camera lens may be moved only once, directly to the desired focus position. Such an approach, however, requires relatively expensive optical components.
Yet another approach determines a focus distance using ultrasonic pulses. For example, Polaroid developed a module that emits ultrasonic pulses and measures the delay of the corresponding return signals. A disadvantage of this approach, however, is that the required hardware is relatively bulky and expensive.
A further approach that may be used involves determining a focus distance using light pulses. For example, surveying equipment emits an infrared laser, receives a corresponding reflection via a photodiode detector, and then determines a distance to the target object based on the received reflection. Using a laser enables measuring long distances with a high degree of accuracy. This approach, however, requires receiving a narrow beam from a target object, and is not effective for a broad-beam, multiple-return path situation, as may be needed for photography.